Farmers are the heart and soul of food security and have been at the forefront of the innovation required to adapt to modern problems something which by-itself serves as an example of a problem and a solution. They have been addressing the new food security issues such as climate change, soil erosion, and sub-optimal resource usage as best they can. One key to addressing these new problems is to accurately predict potential crop yields as such predictive analytics can be used to address in to use predictive analytics to address problems proactively. In this paper, we explain a multi-data source AI application which combines soil nutrient quality, environmental, and remote sensing index data, and cross engineered soil and environmental data with machine learning algorithms to predict the ANN, RF, and fuzzy logic. All data ranked and classified and missing data resolved to the appropriate confidence level and at the appropriate confidence level for data condition. Our proposed model, realized in Java/Weka has produced unparalleled optimum predictive analytics ratings. In the cross engineered environmental data weighted model, banana cultivation was predicted to have 95.83% optimum yield, predicted precision 85.71% with a confidence of 90% and an F1 of 88 predictive F1 to 50.18 tons/ha at a confidence of 98.06%, R². Predictive analytics are self-optimizing and will increase the predictive yield as necessitated that will increase crop yield with decreased added irrigation by remaining crops.
Deore et al. (Wed,) studied this question.